Dissertations / Theses on the topic 'Learning algorithm'
To see the other types of publications on this topic, follow the link: Learning algorithm.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Learning algorithm.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Janagam, Anirudh, and Saddam Hossen. "Analysis of Network Intrusion Detection System with Machine Learning Algorithms (Deep Reinforcement Learning Algorithm)." Thesis, Blekinge Tekniska Högskola, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-17126.
Full text
2
Patel, Darshan D. "Vehicle classification using machine learning algorithm." Thesis, California State University, Long Beach, 2015. http://pqdtopen.proquest.com/#viewpdf?dispub=1604876.
Full textAbstract:
Increasing traffic on roadways requires some real-time system that can collect traffic data and helps us to manage existing road infrastructure. For this purpose, we need a state of art system that can detect and classify vehicles into different categories. We developed an in-node microprocessor-based vehicle classification system to analyze and determine the types of vehicles passing over a 3-axis magnetometer sensor. Our approach for vehicle classification utilizes J48 classification algorithm, which is implemented in machine learning software Weka. J48 is a Quinlan's C4.5 algorithm, an extension of decision tree machine learning based on ID3 algorithm. The decision tree model is generated from a set of features extracted from vehicles passing over the 3-axis sensor. The generated tree model can then be easily implemented on microprocessors. The result of our experiment shows that the vehicle classification system is effective and efficient with the very high accuracy at ~98%.
3
Cui, Yan Hong. "Contributions to statistical machine learning algorithm." Doctoral thesis, University of Cape Town, 2011. http://hdl.handle.net/11427/10284.
Full textAbstract:
This thesis's research focus is on computational statistics along with DEAR (abbreviation of differential equation associated regression) model direction, and that in mind, the journal papers are written as contributions to statistical machine learning algorithm literature.
4
Del, Ben Enrico <1997>. "Reinforcement Learning: a Q-Learning Algorithm for High Frequency Trading." Master's Degree Thesis, Università Ca' Foscari Venezia, 2021. http://hdl.handle.net/10579/20411.
Full textAbstract:
The scope of this work is to test the implementation of an automated trading system based on Reinforcement Learning: a machine learning algorithm in which an intelligent agent acts to maximize its rewards given the environment around it. Indeed, given the environmental inputs and the environmental responses to the actions taken, the agent will learn how to behave in best way possible. In particular, in this work, a Q-Learning algorithm has been used to produce trading signals on the basis of high frequency data of the Limit Order Book for some selected stocks.
5
Cardamone, Dario. "Support Vector Machine a Machine Learning Algorithm." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2017.
Find full textAbstract:
Nella presente tesi di laurea viene preso in considerazione l’algoritmo di classificazione Support Vector Machine. Piu` in particolare si considera la sua formulazione come problema di ottimizazione Mixed Integer Program per la classificazione binaria super- visionata di un set di dati.
6
El-Omari, Jawad A. "Efficient learning methods to tune algorithm parameters." Thesis, University of Warwick, 2013. http://wrap.warwick.ac.uk/58890/.
Full textAbstract:
This thesis focuses on the algorithm configuration problem. In particular, three efficient learning configurators are introduced to tune parameters offline. The first looks into metaoptimization, where the algorithm is expected to solve similar problem instances within varying computational budgets. Standard meta-optimization techniques have to be repeated whenever the available computational budget changes, as the parameters that work well for small budgets, may not be suitable for larger ones. The proposed Flexible Budget method can, in a single run, identify the best parameter setting for all possible computational budgets less than a specified maximum, without compromising solution quality. Hence, a lot of time is saved. This will be shown experimentally. The second regards Racing algorithms which often do not fully utilize the available computational budget to find the best parameter setting, as they may terminate whenever a single parameter remains in the race. The proposed Racing with reset can overcome this issue, and at the same time adapt Racing’s hyper-parameter α online. Experiments will show that such adaptation enables the algorithm to achieve significantly lower failure rates, compared to any fixed α set by the user. The third extends on Racing with reset by allowing it to utilize all the information gathered previously when it adapts α, it also permits Racing algorithms in general to intelligently allocate the budget in each iteration, as opposed to equally allocating it. All developed Racing algorithms are compared to two budget allocators from the Simulation Optimization literature, OCBA and CBA, and to equal allocation to demonstrate under which conditions each performs best in terms of minimizing the probability of incorrect selection.
7
Murphy, Nicholas John. "An online learning algorithm for technical trading." Master's thesis, Faculty of Science, 2019. http://hdl.handle.net/11427/31048.
Full textAbstract:
We use an adversarial expert based online learning algorithm to learn the optimal parameters required to maximise wealth trading zero-cost portfolio strategies. The learning algorithm is used to determine the relative population dynamics of technical trading strategies that can survive historical back-testing as well as form an overall aggregated portfolio trading strategy from the set of underlying trading strategies implemented on daily and intraday Johannesburg Stock Exchange data. The resulting population time-series are investigated using unsupervised learning for dimensionality reduction and visualisation. A key contribution is that the overall aggregated trading strategies are tested for statistical arbitrage using a novel hypothesis test proposed by Jarrow et al. [31] on both daily sampled and intraday time-scales. The (low frequency) daily sampled strategies fail the arbitrage tests after costs, while the (high frequency) intraday sampled strategies are not falsified as statistical arbitrages after costs. The estimates of trading strategy success, cost of trading and slippage are considered along with an offline benchmark portfolio algorithm for performance comparison. In addition, the algorithms generalisation error is analysed by recovering a probability of back-test overfitting estimate using a nonparametric procedure introduced by Bailey et al. [19]. The work aims to explore and better understand the interplay between different technical trading strategies from a data-informed perspective.
8
O'Shea, Timothy James. "Learning from Data in Radio Algorithm Design." Diss., Virginia Tech, 2017. http://hdl.handle.net/10919/89649.
Full textAbstract:
Algorithm design methods for radio communications systems are poised to undergo a massive disruption over the next several years. Today, such algorithms are typically designed manually using compact analytic problem models. However, they are shifting increasingly to machine learning based methods using approximate models with high degrees of freedom, jointly optimized over multiple subsystems, and using real-world data to drive design which may have no simple compact probabilistic analytic form.
Over the past five years, this change has already begun occurring at a rapid pace in several fields. Computer vision tasks led deep learning, demonstrating that low level features and entire end-to-end systems could be learned directly from complex imagery datasets, when a powerful collection of optimization methods, regularization methods, architecture strategies, and efficient implementations were used to train large models with high degrees of freedom.
Within this work, we demonstrate that this same class of end-to-end deep neural network based learning can be adapted effectively for physical layer radio systems in order to optimize for sensing, estimation, and waveform synthesis systems to achieve state of the art levels of performance in numerous applications.
First, we discuss the background and fundamental tools used, then discuss effective strategies and approaches to model design and optimization. Finally, we explore a series of applications across estimation, sensing, and waveform synthesis where we apply this approach to reformulate classical problems and illustrate the value and impact this approach can have on several key radio algorithm design problems.Ph. D.
9
Gunneström, Albert, and Erik Bauer. "Automating dataflow for a machine learning algorithm." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-253068.
Full textAbstract:
Machine learning algorithms can be used to predict the future demand for heat in buildings. This can be used as a decision basis by district heating plants when deciding an appropriate heat output for the plant. This project is based on an existing machine learning model that uses temperature data and the previous heat demand as input data. The model has to be able to make new predictions and display the results continuously in order to be useful for heating plant operators. In this project a program was developed that automatically collects input data, uses this data with the machine learning model and displays the predicted heat demand in a graph. One of the sources for input data does not always provide reliable data and in order to ensure that the program runs continuously and in a robust way, approximations of missing data have to be made. The result is a program that runs continuously but with some constraints on the input data. The input data needs to be able to provide some correct values within the last two days in order for the program run continuously. A comparison between calculated predictions and the actual measured heat demand showed that the predictions were in general higher than the actual values. Some possible causes and solutions were identified but are left for future work.Maskininlärnings-algoritmer kan användas för att göra prediktioner på den framtida efterfrågan på värme i fastigheter. Detta kan användas som ett beslutsunderlag av fjärrvärmeverk för att avgöra en lämplig uteffekt. Detta projektarbete baseras på en befintlig maskininlärnings-modell som använder sig av temperaturdata och tidigare värmedata som inparametrar. Modellen måste kunna göra nya prediktioner och visa resultaten kontinuerligt för att vara användbar för driftpersonal på fjärrvärmeverk. I detta projekt utvecklades ett program som automatiskt samlar in inparameterdata, använder denna data i maskininlärnings-modellen och visar resultaten i en graf. En av källorna för inparameterdata ger inte alltid pålitlig data och för att garantera att programmet körs kontinuerligt och på ett robust vis så måste man approximera inkorrekt data. Resultatet är ett program som kör kontinuerligt men med några restriktioner på inparameterdatan. Inparameterdatan måste ha åtminstone några korrekta värden inom de senaste två dagarna för att programmet ska köras kontinuerligt. En jämförelse mellan beräknade prediktioner och den verkliga uppmätta efterfrågan på värme visade att prediktionerna generellt var högre än de verkliga värdena. Några möjliga orsaker och lösningar identifierades men lämnas till framtida arbeten.
10
Cully, Antoine. "Creative Adaptation through Learning." Thesis, Paris 6, 2015. http://www.theses.fr/2015PA066664/document.
Full textAbstract:
Les robots ont profondément transformé l’industrie manufacturière et sont susceptibles de délivrer de grands bénéfices pour la société, par exemple en intervenant sur des lieux de catastrophes naturelles, lors de secours à la personne ou dans le cadre de la santé et des transports. Ce sont aussi des outils précieux pour la recherche scientifique, comme pour l’exploration des planètes ou des fonds marins. L’un des obstacles majeurs à leur utilisation en dehors des environnements parfaitement contrôlés des usines ou des laboratoires, est leur fragilité. Alors que les animaux peuvent rapidement s’adapter à des blessures, les robots actuels ont des difficultés à faire preuve de créativité lorsqu’ils doivent surmonter un problème inattendu: ils sont limités aux capteurs qu’ils embarquent et ne peuvent diagnostiquer que les situations qui ont été anticipées par leur concepteurs. Dans cette thèse, nous proposons une approche différente qui consiste à laisser le robot apprendre de lui-même un comportement palliant la panne. Cependant, les méthodes actuelles d’apprentissage sont lentes même lorsque l’espace de recherche est petit et contraint. Pour surmonter cette limitation et permettre une adaptation rapide et créative, nous combinons la créativité des algorithmes évolutionnistes avec la rapidité des algorithmes de recherche de politique à travers trois contributions : les répertoires comportementaux, l’adaptation aux dommages et le transfert de connaissance entre plusieurs tâches. D’une manière générale, ces travaux visent à apporter les fondations algorithmiques permettant aux robots physiques d’être plus robustes, performants et autonomesRobots have transformed many industries, most notably manufacturing, and have the power to deliver tremendous benefits to society, for example in search and rescue, disaster response, health care, and transportation. They are also invaluable tools for scientific exploration of distant planets or deep oceans. A major obstacle to their widespread adoption in more complex environments and outside of factories is their fragility. While animals can quickly adapt to injuries, current robots cannot “think outside the box” to find a compensatory behavior when they are damaged: they are limited to their pre-specified self-sensing abilities, which can diagnose only anticipated failure modes and strongly increase the overall complexity of the robot. In this thesis, we propose a different approach that considers having robots learn appropriate behaviors in response to damage. However, current learning techniques are slow even with small, constrained search spaces. To allow fast and creative adaptation, we combine the creativity of evolutionary algorithms with the learning speed of policy search algorithms through three contributions: the behavioral repertoires, the damage recovery using these repertoires and the transfer of knowledge across tasks. Globally, this work aims to provide the algorithmic foundations that will allow physical robots to be more robust, effective and autonomous
11
Ahmad, Jamil. "A novel learning algorithm for feedforward neural network." Thesis, King's College London (University of London), 1994. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.404310.
Full text
12
Gaudreau, Balderrama Amanda Dawn. "Advanced therapy learning algorithm for spinal cord stimulation." Thesis, Massachusetts Institute of Technology, 2010. http://hdl.handle.net/1721.1/62639.
Full textAbstract:
Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2010.Cataloged from PDF version of thesis.
Includes bibliographical references (p. 85-87).
Spinal Cord Stimulation (SCS) is a technique used to treat chronic pain and has been shown to be an effective method of treatment, both financially and socioeconomically. Stimulating electrodes are surgically implanted into the epidural space, outside the dura, a protective sac filled with cerebral spinal fluid (CSF) surrounding the spinal cord. The thickness of the CSF changes according to body orientation, causing the distance between the stimulating electrodes and the spinal cord to vary. This phenomenon has been reported to cause painful or ineffective stimulation. In order to detect postural behavior and adjust SCS parameters accordingly, a tri-axial accelerometer based algorithm has been developed. The algorithm enables patients to adjust stimulation therapy parameters real-time, associates the patient indicated parameters with a vector, and stores them in a therapy library. Stimulation therapy parameters are then automatically selected by classifying incoming TA data according to the vectors in the therapy library, providing individualized, closed-loop stimulation therapy.
by Amanda Dawn Gaudreau Balderrama.
M.Eng.
13
Cai, Zhonglun. "Iterative learning control : algorithm development and experimental benchmarking." Thesis, University of Southampton, 2009. https://eprints.soton.ac.uk/66415/.
Full textAbstract:
This thesis concerns the general area of experimental benchmarking of Iterative Learning Control (ILC) algorithms using two experimental facilities. ILC is an approach which is suitable for applications where the same task is executed repeatedly over the necessarily finite time duration, known as the trial length. The process is reset prior to the commencement of each execution. The basic idea of ILC is to use information from previously executed trials to update the control input to be applied during the next one. The first experimental facility is a non-minimum phase electro-mechanical system and the other is a gantry robot whose basic task is to pick and place objects on a moving conveyor under synchronization and in a fixed finite time duration that replicates many tasks encountered in the process industries. Novel contributions are made in both the development of new algorithms and, especially, in the analysis of experimental results both of a single algorithm alone and also in the comparison of the relative performance of different algorithms. In the case of non-minimum phase systems, a new algorithm, named Reference Shift ILC (RSILC) is developed that is of a two loop structure. One learning loop addresses the system lag and another tackles the possibility of a large initial plant input commonly encountered when using basic iterative learning control algorithms. After basic algorithm development and simulation studies, experimental results are given to conclude that performance improvement over previously reported algorithms is reasonable. The gantry robot has been previously used to experimentally benchmark a range of simple structure ILC algorithms, such as those based on the ILC versions of the classical proportional plus derivative error actuated controllers, and some state-space based optimal ILC algorithms. Here these results are extended by the first ever detailed experimental study of the performance of stochastic ILC algorithms together with some modifications necessary to their configuration in order to increase performance. The majority of the currently reported ILC algorithms mainly focus on reducing the trial-to-trial error but it is known that this may come at the cost of poor or unacceptable performance along the trial dynamics. Control theory for discrete linear repetitive processes is used to design ILC control laws that enable the control of both trial-to-trial error convergence and along the trial dynamics. These algorithms can be computed using Linear Matrix Inequalities (LMIs) and again the results of experimental implementation on the gantry robot are given. These results are the first ever in this key area and represent a benchmark against which alternatives can be compared. In the concluding chapter, a critical overview of the results presented is given together with areas for both short and medium term further research.
14
Olsson, Rasmus, and Jens Egeland. "Reinforcement Learning Routing Algorithm for Bluetooth Mesh Networks." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-234287.
Full textAbstract:
Today’s office and home environments are moving towards more connected dig- ital infrastructures, meaning there are multiple heterogeneous devices that uses short-range communication to stay connected. Mobile phones, tablets, lap- tops, sensors, printers are examples of devices in such environments. From this, the Internet of Things (IoT) paradigm arises, and to enable it, energy efficient machine-to-machine (M2M) communications are needed. Our study will use Bluetooth Low Energy (BLE) technology for communication between devices, and it demonstrates the impact of routing algorithms in such networks. With the goal to increase the network lifetime, a distributed and dynamic Reinforce- ment Learning (RL) routing algorithm is proposed. The algorithm is based on a RL technique called Q-learning. Performance analysis is performed in different scenarios comparing the proposed algorithm against two static and centralized reference routing algorithms. The results show that our proposed RL routing algorithm performs better as the node degree of the topology increases. Com- pared to the reference algorithms the proposed algorithm can handle a higher load on the network with significant performance improvement, due to the dy- namic change of routes. The increase in network lifetime with 75 devices is 124% and 100 devices is 349%, because of the ability to change routes as time passes which is emphasized when the node degree increases. For 35, 55 and 75 devices the average node degrees are 2.21, 2.39 and 2.54. On a lower number of devices our RL routing algorithm performs nearly as good as the best refer- ence algorithm, the Energy Aware Routing (EAR) algorithm, with a decrease in network lifetime around 19% on 35 devices and 10% on 55 devices. A decrease in the network lifetime on lower number of devices is because of the cost for learning new paths is higher than the gain from exploring multiple paths.Dagens kontors- och hemmiljöer rör sig mot mer sammankopplad digital in-frastruktur, vilket innebär att det finns många heterogena enheter som behöver kommunicera med varandra på korta avstånd. Mobiltelefoner, tablets, bärbara datorer, sensorer, skrivare är exempel på enheter i sådana miljöer. Utifrån detta uppkommer IoT, och för att möjliggöra det, behövs energieffektiva M2M kom-munikationslösningar. Vår studie kommer att anvanda BLE teknik för kommu-nikation mellan enheter, och den kommer att demonstrera effekterna av routing algoritmer i sådana nätverk. Med målet att öka livstiden för nätverket föreslås en distribuerad och dynamisk RL routing algoritm baserad på Q-learning. En jämförelse mellan den föreslagna algoritmen och de två statiska och centraliser-ade referensalgoritmerna görs i olika simulerings scenarier. Resultaten visar att vår föreslagna RL routing algoritm fungerar bättre när nod graden i topologin ökar. Jämfört med referensalgoritmerna kan den föreslagna algoritmen hantera en högre belastning på nätverket med betydande prestandaförbättring, tack vare den dynamiska förändringen av rutter som leder till en bättre belastningsbal-ans. Ökningen i nätverkslivstiden med 75 enheter är 124% och med 100 enheter är ökningen 349%, på grund av förmågan att byta rutter vilket syns tydligare när nodgraden ökar. För 35, 55 och 75 enheter är nodgraderna 2.21, 2.39 och 2.54. Vid ett lägre antal enheter presterar vår RL routing algoritm nästan lika bra som den bästa referensalgoritmen, EAR, med en minskning av nätverks livstiden på runt 19% med 35 enheter och 10% med 55 enheter. En minskning av nätverks livstiden på lägre antal enheter beror på att kostnaden för att lära sig nya vägar är högre än vinsten från att utforska flera vägar.
15
Gondlyala, Siddharth Rao. "Enhancing the JPEG Ghost Algorithm using Machine Learning." Thesis, Blekinge Tekniska Högskola, Institutionen för datavetenskap, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-20692.
Full textAbstract:
Background: With the boom in the internet space and social media platforms, a large number of images are being shared. With this rise and advancements in technology, many image editing tools have made their way to giving rise to digital image manipulation. Being able to differentiate a forged image is vital to avoid misinformation or misrepresentation. This study focuses on the splicing image forgery to localizes the forged region in the tampered image. Objectives: The main purpose of the thesis is to extend the capability of the JPEG Ghost model by localizing the tampering in the image. This is done by analyzing the difference curves formed by compressions in the tampered image, and thereafter comparing the performance of the models. Methods: The study is carried out by two research methods; one being a Literature Review, whose main goal is gaining insights on the existing studies in terms of the approaches and techniques followed; and the second being Experiment; whose main goal is to improve the JPEG ghost algorithm by localizing the forged area in a tampered image and to compare three machine learning models based on the performance metrics. The machine learning models that are compared are Random Forest, XGBoost, and Support Vector Machine. Results: The performance of the above-mentioned models has been compared with each other on the same dataset. Results from the experiment showed that XGBoost had the best overall performance over other models with the Jaccard Index value of 79.8%. Conclusions: The research revolves around localization of the forged region in a tampered image using the concept of JPEG ghosts. This is We have concluded that the performance of XGBoost model is the best, followed by Random Forest and then Support Vector Machine.
16
Zhang, Yi. "Groupwise Distance Learning Algorithm for User Recommendation Systems." University of Cincinnati / OhioLINK, 2016. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1471347509.
Full text
17
Shao, Yunming. "Image-based Perceptual Learning Algorithm for Autonomous Driving." The Ohio State University, 2017. http://rave.ohiolink.edu/etdc/view?acc_num=osu1503302777088283.
Full text
18
Lowton, Andrew D. "A constructive learning algorithm based on back-propagation." Thesis, Aston University, 1995. http://publications.aston.ac.uk/10663/.
Full textAbstract:
There are been a resurgence of interest in the neural networks field in recent years, provoked in part by the discovery of the properties of multi-layer networks. This interest has in turn raised questions about the possibility of making neural network behaviour more adaptive by automating some of the processes involved. Prior to these particular questions, the process of determining the parameters and network architecture required to solve a given problem had been a time consuming activity. A number of researchers have attempted to address these issues by automating these processes, concentrating in particular on the dynamic selection of an appropriate network architecture. The work presented here specifically explores the area of automatic architecture selection; it focuses upon the design and implementation of a dynamic algorithm based on the Back-Propagation learning algorithm. The algorithm constructs a single hidden layer as the learning process proceeds using individual pattern error as the basis of unit insertion. This algorithm is applied to several problems of differing type and complexity and is found to produce near minimal architectures that are shown to have a high level of generalisation ability. (DX 187, 339)
19
Cully, Antoine. "Creative Adaptation through Learning." Electronic Thesis or Diss., Paris 6, 2015. http://www.theses.fr/2015PA066664.
Full textAbstract:
Les robots ont profondément transformé l’industrie manufacturière et sont susceptibles de délivrer de grands bénéfices pour la société, par exemple en intervenant sur des lieux de catastrophes naturelles, lors de secours à la personne ou dans le cadre de la santé et des transports. Ce sont aussi des outils précieux pour la recherche scientifique, comme pour l’exploration des planètes ou des fonds marins. L’un des obstacles majeurs à leur utilisation en dehors des environnements parfaitement contrôlés des usines ou des laboratoires, est leur fragilité. Alors que les animaux peuvent rapidement s’adapter à des blessures, les robots actuels ont des difficultés à faire preuve de créativité lorsqu’ils doivent surmonter un problème inattendu: ils sont limités aux capteurs qu’ils embarquent et ne peuvent diagnostiquer que les situations qui ont été anticipées par leur concepteurs. Dans cette thèse, nous proposons une approche différente qui consiste à laisser le robot apprendre de lui-même un comportement palliant la panne. Cependant, les méthodes actuelles d’apprentissage sont lentes même lorsque l’espace de recherche est petit et contraint. Pour surmonter cette limitation et permettre une adaptation rapide et créative, nous combinons la créativité des algorithmes évolutionnistes avec la rapidité des algorithmes de recherche de politique à travers trois contributions : les répertoires comportementaux, l’adaptation aux dommages et le transfert de connaissance entre plusieurs tâches. D’une manière générale, ces travaux visent à apporter les fondations algorithmiques permettant aux robots physiques d’être plus robustes, performants et autonomesRobots have transformed many industries, most notably manufacturing, and have the power to deliver tremendous benefits to society, for example in search and rescue, disaster response, health care, and transportation. They are also invaluable tools for scientific exploration of distant planets or deep oceans. A major obstacle to their widespread adoption in more complex environments and outside of factories is their fragility. While animals can quickly adapt to injuries, current robots cannot “think outside the box” to find a compensatory behavior when they are damaged: they are limited to their pre-specified self-sensing abilities, which can diagnose only anticipated failure modes and strongly increase the overall complexity of the robot. In this thesis, we propose a different approach that considers having robots learn appropriate behaviors in response to damage. However, current learning techniques are slow even with small, constrained search spaces. To allow fast and creative adaptation, we combine the creativity of evolutionary algorithms with the learning speed of policy search algorithms through three contributions: the behavioral repertoires, the damage recovery using these repertoires and the transfer of knowledge across tasks. Globally, this work aims to provide the algorithmic foundations that will allow physical robots to be more robust, effective and autonomous
20
Shi, Haijian. "Best-first Decision Tree Learning." The University of Waikato, 2007. http://hdl.handle.net/10289/2317.
Full textAbstract:
In best-first top-down induction of decision trees, the best split is added in each step (e.g. the split that maximally reduces the Gini index). This is in contrast to the standard depth-first traversal of a tree. The resulting tree will be the same, just how it is built is different. The objective of this project is to investigate whether it is possible to determine an appropriate tree size on practical datasets by combining best-first decision tree growth with cross-validation-based selection of the number of expansions that are performed. Pre-pruning, post-pruning, CART-pruning can be performed this way to compare.
21
Rong, Ruichen. "Developing a Phylogeny Based Machine Learning Algorithm for Metagenomics." Thesis, University of North Texas, 2017. https://digital.library.unt.edu/ark:/67531/metadc1011752/.
Full textAbstract:
Metagenomics is the study of the totality of the complete genetic elements discovered from a defined environment. Different from traditional microbiology study, which only analyzes a small percent of microbes that could survive in laboratory, metagenomics allows researchers to get entire genetic information from all the samples in the communities. So metagenomics enables understanding of the target environments and the hidden relationships between bacteria and diseases. In order to efficiently analyze the metagenomics data, cutting-edge technologies for analyzing the relationships among microbes and communities are required. To overcome the challenges brought by rapid growth in metagenomics datasets, advances in novel methodologies for interpreting metagenomics data are clearly needed.
The first two chapters of this dissertation summarize and compare the widely-used methods in metagenomics and integrate these methods into pipelines. Properly analyzing metagenomics data requires a variety of bioinformatcis and statistical approaches to deal with different situations. The raw reads from sequencing centers need to be processed and denoised by several steps and then be further interpreted by ecological and statistical analysis. So understanding these algorithms and combining different approaches could potentially reduce the influence of noises and biases at different steps. And an efficient and accurate pipeline is important to robustly decipher the differences and functionality of bacteria in communities.
Traditional statistical analysis and machine learning algorithms have their limitations on analyzing metagenomics data. Thus, rest three chapters describe a new phylogeny based machine learning and feature selection algorithm to overcome these problems. The new method outperforms traditional algorithms and can provide more robust candidate microbes for further analysis. With the frowing sample size, deep neural network could potentially describe more complicated characteristic of data and thus improve model accuracy. So a deep learning framework is designed on top of the shallow learning algorithm stated above in order to further improve the prediction and selection accuracy.
The present dissertation work provides a powerful tool that utilizes machine learning techniques to identify signature bacteria and key information from huge amount of metagenomics data.
22
Dimitriadou, Evgenia, Andreas Weingessel, and Kurt Hornik. "A voting-merging clustering algorithm." SFB Adaptive Information Systems and Modelling in Economics and Management Science, WU Vienna University of Economics and Business, 1999. http://epub.wu.ac.at/94/1/document.pdf.
Full textAbstract:
In this paper we propose an unsupervised voting-merging scheme that is capable of clustering data sets, and also of finding the number of clusters existing in them. The voting part of the algorithm allows us to combine several runs of clustering algorithms resulting in a common partition. This helps us to overcome instabilities of the clustering algorithms and to improve the ability to find structures in a data set. Moreover, we develop a strategy to understand, analyze and interpret these results. In the second part of the scheme, a merging procedure starts on the clusters resulting by voting, in order to find the number of clusters in the data set.Series: Working Papers SFB "Adaptive Information Systems and Modelling in Economics and Management Science"
23
Topalli, Ayca Kumluca. "Hybrid Learning Algorithm For Intelligent Short-term Load Forecasting." Phd thesis, METU, 2003. http://etd.lib.metu.edu.tr/upload/627505/index.pdf.
Full textAbstract:
Short-term load forecasting (STLF) is an important part of the power generation
process. For years, it has been achieved by traditional approaches stochastic like
time seriesbut, new methods based on artificial intelligence emerged recently in literature and started to replace the old ones in the industry. In order to follow the latest developments and to have a modern system, it is aimed to make a research on STLF in Turkey, by neural networks. For this purpose, a method is proposed to forecast Turkey&rsquo
s total electric load one day in advance. A hybrid learning scheme that combines off-line learning with real-time forecasting is developed to make use of the available past data for adapting the weights and to further adjust these connections according to the changing conditions. It is also suggested to tune the step size iteratively for better accuracy. Since a single neural network model cannot cover all load types, data are clustered due to the differences in their characteristics. Apart from this, special days are extracted from the normal training sets and handled separately. In this way, a solution is proposed for all load types, including working days, weekends and special holidays. For the selection of input parameters, a technique based on principal component analysis is suggested. A traditional ARMA model is constructed for the same data as a benchmark and results are compared. Proposed method gives lower percent errors all the time, especially for holiday loads. The average error for year 2002 data is obtained as 1.60%.
24
Ghosh, Ranadhir, and n/a. "A Novel Hybrid Learning Algorithm For Artificial Neural Networks." Griffith University. School of Information Technology, 2003. http://www4.gu.edu.au:8080/adt-root/public/adt-QGU20030808.162355.
Full textAbstract:
Last few decades have witnessed the use of artificial neural networks (ANN) in many real-world applications and have offered an attractive paradigm for a broad range of adaptive complex systems. In recent years ANN have enjoyed a great deal of success and have proven useful in wide variety pattern recognition or feature extraction tasks. Examples include optical character recognition, speech recognition and adaptive control to name a few. To keep the pace with its huge demand in diversified application areas, many different kinds of ANN architecture and learning types have been proposed by the researchers to meet varying needs. A novel hybrid learning approach for the training of a feed-forward ANN has been proposed in this thesis. The approach combines evolutionary algorithms with matrix solution methods such as singular value decomposition, Gram-Schmidt etc., to achieve optimum weights for hidden and output layers. The proposed hybrid method is to apply evolutionary algorithm in the first layer and least square method (LS) in the second layer of the ANN. The methodology also finds optimum number of hidden neurons using a hierarchical combination methodology structure for weights and architecture. A learning algorithm has many facets that can make a learning algorithm good for a particular application area. Often there are trade offs between classification accuracy and time complexity, nevertheless, the problem of memory complexity remains. This research explores all the different facets of the proposed new algorithm in terms of classification accuracy, convergence property, generalization ability, time and memory complexity.
25
Liang, Aileen H. "Rough set-based distance learning algorithm and its implementation." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 2001. http://www.collectionscanada.ca/obj/s4/f2/dsk3/ftp04/MQ60237.pdf.
Full text
26
Köpf, Christian Rudolf. "Meta-learning: strategies, implementations, and evaluations for algorithm selection /." Berlin : Aka, 2006. http://deposit.ddb.de/cgi-bin/dokserv?id=2745748&prov=M&dok_var=1&dok_ext=htm.
Full text
27
Hirotsu, Kenichi. "Neural network hardware with random weight change learning algorithm." Diss., Georgia Institute of Technology, 1993. http://hdl.handle.net/1853/15765.
Full text
28
Laflamme, Simon M. Eng Massachusetts Institute of Technology. "Online learning algorithm for structural control using magnetorheological actuators." Thesis, Massachusetts Institute of Technology, 2007. http://hdl.handle.net/1721.1/39271.
Full textAbstract:
Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Civil and Environmental Engineering, 2007.Includes bibliographical references (p. 83-84).
Magnetorheological actuators are promising devices for mitigating vibrations because they only require a fraction of energy for a similar performance to active control. Conversely, these semi-active devices have limited maximum forces and are hard to model due to the rheological properties of their fluid. When considering structural control, classical theories necessitate full knowledge of the structural dynamic states and properties most of which can only be estimated when considering large-scale control, which may be difficult or inaccurate for complicated geometries due to the non-linear behaviour of structures. Additionally, most of these theories do not take into account the response delay of the actuators which may result in structural instabilities. To address the problem, learning algorithms using offline learning have been proposed in order to have the structure learn its behaviour, but they can be perceived as unrealistic because earthquake data can hardly be produced to train these schemes. Here, an algorithm using online learning feedback is proposed to address this problem where the structure observes, compares and adapts its performance at each time step, analogous to a child learning his or her motor functions.
(cont.) The algorithm uses a machine learning technique, Gaussian kernels, to prescribe forces upon structural states, where states are evaluated strictly based on displacement and acceleration feedback. The algorithm has been simulated and performances assessed by comparing it with two classical control theories: clipped-optimal and passive controls. The proposed scheme is found to be stable and performs well in mitigating vibrations for a low energy input, but does not perform as well compared to clipped-optimal case. This relative performance would be expected to be better for large-scale structures because of the adaptability of the proposed algorithm.
by Simon Laflamme.
M.Eng.
29
Wang, Grant J. (Grant Jenhorn) 1979. "A special algorithm for learning mixtures of spherical Gaussians." Thesis, Massachusetts Institute of Technology, 2003. http://hdl.handle.net/1721.1/87899.
Full text
30
Gandhi, Sachin. "Learning from a Genetic Algorithm with Inductive Logic Programming." Ohio University / OhioLINK, 2005. http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1125511501.
Full text
31
Ghosh, Ranadhir. "A Novel Hybrid Learning Algorithm For Artificial Neural Networks." Thesis, Griffith University, 2003. http://hdl.handle.net/10072/365961.
Full textAbstract:
Last few decades have witnessed the use of artificial neural networks (ANN) in many real-world applications and have offered an attractive paradigm for a broad range of adaptive complex systems. In recent years ANN have enjoyed a great deal of success and have proven useful in wide variety pattern recognition or feature extraction tasks. Examples include optical character recognition, speech recognition and adaptive control to name a few. To keep the pace with its huge demand in diversified application areas, many different kinds of ANN architecture and learning types have been proposed by the researchers to meet varying needs. A novel hybrid learning approach for the training of a feed-forward ANN has been proposed in this thesis. The approach combines evolutionary algorithms with matrix solution methods such as singular value decomposition, Gram-Schmidt etc., to achieve optimum weights for hidden and output layers. The proposed hybrid method is to apply evolutionary algorithm in the first layer and least square method (LS) in the second layer of the ANN. The methodology also finds optimum number of hidden neurons using a hierarchical combination methodology structure for weights and architecture. A learning algorithm has many facets that can make a learning algorithm good for a particular application area. Often there are trade offs between classification accuracy and time complexity, nevertheless, the problem of memory complexity remains. This research explores all the different facets of the proposed new algorithm in terms of classification accuracy, convergence property, generalization ability, time and memory complexity.Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Information Technology
Full Text
32
Dahlberg, Love. "Dynamic algorithm selection for machine learning on time series." Thesis, Karlstads universitet, Institutionen för matematik och datavetenskap (from 2013), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kau:diva-72576.
Full textAbstract:
We present a software that can dynamically determine what machine learning algorithm is best to use in a certain situation given predefined traits. The produced software uses ideal conditions to exemplify how such a solution could function. The software is designed to train a selection algorithm that can predict the behavior of the specified testing algorithms to derive which among them is the best. The software is used to summarize and evaluate a collection of selection algorithm predictions to determine which testing algorithm was the best during that entire period. The goal of this project is to provide a prediction evaluation software solution can lead towards a realistic implementation.
33
Cook, Philip R. "Limitations and Extensions of the WoLF-PHC Algorithm." Diss., CLICK HERE for online access, 2007. http://contentdm.lib.byu.edu/ETD/image/etd2109.pdf.
Full text
34
Dash, Sajal. "Exploring the Landscape of Big Data Analytics Through Domain-Aware Algorithm Design." Diss., Virginia Tech, 2020. http://hdl.handle.net/10919/99798.
Full textAbstract:
Experimental and observational data emerging from various scientific domains necessitate
fast, accurate, and low-cost analysis of the data. While exploring the landscape
of big data analytics, multiple challenges arise from three characteristics of big data:
the volume, the variety, and the velocity. High volume and velocity of the data warrant
a large amount of storage, memory, and compute power while a large variety
of data demands cognition across domains. Addressing domain-intrinsic properties of
data can help us analyze the data efficiently through the frugal use of high-performance
computing (HPC) resources. In this thesis, we present our exploration of the data analytics
landscape with domain-aware approximate and incremental algorithm design.
We propose three guidelines targeting three properties of big data for domain-aware
big data analytics: (1) explore geometric and domain-specific properties of high dimensional
data for succinct representation, which addresses the volume property, (2) design
domain-aware algorithms through mapping of domain problems to computational problems,
which addresses the variety property, and (3) leverage incremental arrival of data
through incremental analysis and invention of problem-specific merging methodologies,
which addresses the velocity property. We demonstrate these three guidelines through
the solution approaches of three representative domain problems.
We present Claret, a fast and portable parallel weighted multi-dimensional scaling
(WMDS) tool, to demonstrate the application of the first guideline. It combines algorithmic
concepts extended from the stochastic force-based multi-dimensional scaling
(SF-MDS) and Glimmer. Claret computes approximate weighted Euclidean distances
by combining a novel data mapping called stretching and Johnson Lindestrauss' lemma
to reduce the complexity of WMDS from O(f(n)d) to O(f(n) log d). In demonstrating
the second guideline, we map the problem of identifying multi-hit combinations
of genetic mutations responsible for cancers to weighted set cover (WSC) problem by
leveraging the semantics of cancer genomic data obtained from cancer biology. Solving
the mapped WSC with an approximate algorithm, we identified a set of multi-hit
combinations that differentiate between tumor and normal tissue samples. To identify
three- and four-hits, which require orders of magnitude larger computational power, we
have scaled out the WSC algorithm on a hundred nodes of Summit supercomputer. In
demonstrating the third guideline, we developed a tool iBLAST to perform an incremental
sequence similarity search. Developing new statistics to combine search results
over time makes incremental analysis feasible. iBLAST performs (1+δ)/δ times faster
than NCBI BLAST, where δ represents the fraction of database growth. We also explored
various approaches to mitigate catastrophic forgetting in incremental training
of deep learning models.Doctor of Philosophy
Experimental and observational data emerging from various scientific domains necessitate fast, accurate, and low-cost analysis of the data. While exploring the landscape of big data analytics, multiple challenges arise from three characteristics of big data: the volume, the variety, and the velocity. Here volume represents the data's size, variety represents various sources and formats of the data, and velocity represents the data arrival rate. High volume and velocity of the data warrant a large amount of storage, memory, and computational power. In contrast, a large variety of data demands cognition across domains. Addressing domain-intrinsic properties of data can help us analyze the data efficiently through the frugal use of high-performance computing (HPC) resources. This thesis presents our exploration of the data analytics landscape with domain-aware approximate and incremental algorithm design. We propose three guidelines targeting three properties of big data for domain-aware big data analytics: (1) explore geometric (pair-wise distance and distribution-related) and domain-specific properties of high dimensional data for succinct representation, which addresses the volume property, (2) design domain-aware algorithms through mapping of domain problems to computational problems, which addresses the variety property, and (3) leverage incremental data arrival through incremental analysis and invention of problem-specific merging methodologies, which addresses the velocity property. We demonstrate these three guidelines through the solution approaches of three representative domain problems. We demonstrate the application of the first guideline through the design and development of Claret. Claret is a fast and portable parallel weighted multi-dimensional scaling (WMDS) tool that can reduce the dimension of high-dimensional data points. In demonstrating the second guideline, we identify combinations of cancer-causing gene mutations by mapping the problem to a well known computational problem known as the weighted set cover (WSC) problem. We have scaled out the WSC algorithm on a hundred nodes of Summit supercomputer to solve the problem in less than two hours instead of an estimated hundred years. In demonstrating the third guideline, we developed a tool iBLAST to perform an incremental sequence similarity search. This analysis was made possible by developing new statistics to combine search results over time. We also explored various approaches to mitigate the catastrophic forgetting of deep learning models, where a model forgets to perform machine learning tasks efficiently on older data in a streaming setting.
35
Vaaler, Erik Garth. "A machine learning based logic branching algorithm for automated assembly." Thesis, Massachusetts Institute of Technology, 1991. http://hdl.handle.net/1721.1/40555.
Full textAbstract:
Thesis (Sc. D.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering, 1991.Title as it appears in the M.I.T. Graduate List, Feb. 1991: A logic branching based machine learning algorithm for automated assembly.
Includes bibliographical references (p. 95-100).
by Erik Garth Vaaler.
Sc.D.
36
Ho, Chang-An, and 何長安. "Safe Reinforcement Learning based Sequential Perturbation Learning Algorithm." Thesis, 2009. http://ndltd.ncl.edu.tw/handle/63234750154932788712.
Full textAbstract:
碩士國立交通大學
電機與控制工程系所
97
This article is about sequential perturbation learning architecture through safe reinforcement learning (SRL-SP) which based on the concept of linear search to apply perturbations on each weight value of the neural network. The evaluation of value of function between pre-perturb and post-perturb network is executed after the perturbations are applied, so as to update the weights. Applying perturbations can avoid the solution form the phenomenon which falls into the hands of local solution and oscillating in the solution space that decreases the learning efficiency. Besides, in the reinforcement learning structure, use the Lyapunov design methods to set the learning objective and pre-defined set of the goal state. This method would greatly reduces the learning time, in other words, it can rapidly guide the plant’s state into the goal state. During the simulation, use the n-mass inverted pendulum model to perform the experiment of humanoid robot model. To prove the method in this article is more effective in learning.
37
Gorgadze, Luka. "Flipped classes for algorithm learning." Master's thesis, 2016. http://hdl.handle.net/10198/13671.
Full textAbstract:
Recent years have brought the need for new pedagogical approaches, that appeal to the involvement and participation of students in the learning process. One of this approaches is the flipped classroom, which gives to students the possibility to prepare for the next class, through pre-recorded video lectures and close-ended problems. There is some research going on around this model of teaching-learning, showing some promising results.
The purpose of this work is to contribute to this line of research by designing and applying an experiment to compare the efficiency of the flipped classes methodology to the traditional classes. The research hypothesis is that the flipped classes methodology is an efficient method for teaching algorithms.
The results were positive, although mildly. The reason for this was the fact that the experiment lasted only two weeks, not giving enough time for students to fully understand how to learn by the flipped classes and to get used to it. Thus, longer experiments are recommended in order to check full capacity of this method.Nos últimos anos, tem-se vindo a assistir ao aparecimento de várias técnicas e abordagens pedagógicas que procuram incrementar o papel ativo dos alunos no próprio processo de aprendizagem. Uma das abordagens, designada por aula invertida (do inglês flipped classes), incentiva o aluno a preparar a aula antecipadamente, por intermédio de vídeos, conteúdo teórico e problemas para resolver. Há alguma investigação em torno desta abordagem, com resultados positivos. O objetivo deste trabalho é contribuir para a investigação deste tipo de abordagem, investigando trabalho relacionado e comparando a eficiência das aulas invertidas com as aulas tradicionais. A hipótese colocada é que as aulas invertidas constituem um método eficiente para a aprendizagem de algoritmia. Os resultados obtidos foram positivos, o que confirma a recomendação feita em certos trabalhos relacionados no sentido de adotar aulas invertidas em algumas áreas. No entanto, os resultados não assinalam uma diferença considerável com as técnicas tradicionais, provavelmente devido ao facto de a experiência decorrer durante duas semanas apenas, não dando tempo suficiente para os alunos se ambientarem e incorporarem a filosofia desta abordagem.
ბოლო წლებში განსაკუთრებით გამოიკვეთა ისეთი პედაგოგიური მოდელების არსებობის და გამოყენების საჭიროება, რომლებიც განაპირობებენ სტუდენტის აქტიურ ჩართულობას სასწავლო პროცესში. ერთ-ერთი ასეთი მოდელია „შებრუნებული საკლასო ოთახი“ (The Flipped Classroom), რომელიც საშუალებას აძლევს სტუდენტებს მოემზადონ ყოველი შემდეგი ლექციისთვის წინასწარ, ვიდეო ლექციის და სავარჯიშოების საშუალებით. ამ მოდელის გარშემო არაერთი კვლევა ჩატარდა ბოლო ხანებში, რომლებიც ძირითადად დადებითად აფასებენ მას. ამ ნაშრომის მიზანია გარკვეული წვლილის შეტანა შებრუნებული საკლასო ოთახის გარშემო მიმდინარე კვლევაში ექსპერიმენტის შექმნით და ჩატარებით, რომელიც შეადარებს ერთმანეთს შებრუნებული და ტრადიციული სწავლების მოდელების ეფექტურობას. კვლევის ჰიპოთეზა მდგომარეობს შემდეგში, შებრუნებული საკლასო ოთახი ეფექტური მოდელია ალგორითმების სწავლებისათვის. კვლევამ პოზიტიური შედეგები აჩვენა, თუმცა არცისე მკვეთრი. ამის მიზეზი ისაა, რომ ექსპერიმენტის ხანგრძლივობა მხოლოდ ორი კვირა იყო. რაც არ აღმოჩნდა საკმარისი იმისათვის რომ სტუდენტებს აეთვისებინად ის თუ როგორ უნდა ისწავლონ შებრუნებული საკლასო ოთახის მეშვეობით. აქედან გამომდინარე ამ სწავლების მეთოდის სრული პოტენციალის დასადგენად რეკომენდირებულია უფრო ხანგრძლივი ექსპერიმენტების ჩატარება.
38
Hurlbert, Anya, and Tomaso Poggio. "Learning a Color Algorithm from Examples." 1987. http://hdl.handle.net/1721.1/5601.
Full textAbstract:
We show that a color algorithm capable of separating illumination from reflectance in a Mondrian world can be learned from a set of examples. The learned algorithm is equivalent to filtering the image data---in which reflectance and illumination are mixed---through a center-surround receptive field in individual chromatic channels. The operation resembles the "retinex" algorithm recently proposed by Edwin Land. This result is a specific instance of our earlier results that a standard regularization algorithm can be learned from examples. It illustrates that the natural constraints needed to solve a problemsin inverse optics can be extracted directly from a sufficient set of input data and the corresponding solutions. The learning procedure has been implemented as a parallel algorithm on the Connection Machine System.
39
Lakshmanan, K. "Online Learning and Simulation Based Algorithms for Stochastic Optimization." Thesis, 2012. http://etd.iisc.ac.in/handle/2005/3245.
Full textAbstract:
In many optimization problems, the relationship between the objective and parameters is not known. The objective function itself may be stochastic such as a long-run average over some random cost samples. In such cases finding the gradient of the objective is not possible. It is in this setting that stochastic approximation algorithms are used. These algorithms use some estimates of the gradient and are stochastic in nature. Amongst gradient estimation techniques, Simultaneous Perturbation Stochastic Approximation (SPSA) and Smoothed Functional(SF) scheme are widely used. In this thesis we have proposed a novel multi-time scale quasi-Newton based smoothed functional (QN-SF) algorithm for unconstrained as well as constrained optimization. The algorithm uses the smoothed functional scheme for estimating the gradient and the quasi-Newton method to solve the optimization problem. The algorithm is shown to converge with probability one.
We have also provided here experimental results on the problem of optimal routing in a multi-stage network of queues. Policies like Join the Shortest Queue or Least Work Left assume knowledge of the queue length values that can change rapidly or hard to estimate. If the only information available is the expected end-to-end delay as with our case, such policies cannot be used. The QN-SF based probabilistic routing algorithm uses only the total end-to-end delay for tuning the probabilities. We observe from the experiments that the QN-SF algorithm has better performance than the gradient and Jacobi versions of Newton based smoothed functional algorithms. Next we consider constrained routing in a similar queueing network. We extend the QN-SF algorithm to this case. We study the convergence behavior of the algorithm and observe that the constraints are satisfied at the point of convergence. We provide experimental results for the constrained routing setup as well.
Next we study reinforcement learning algorithms which are useful for solving Markov Decision Process(MDP) when the precise information on transition probabilities is not known. When the state, and action sets are very large, it is not possible to store all the state-action tuples. In such cases, function approximators like neural networks have been used. The popular Q-learning algorithm is known to diverge when used with linear function approximation due to the ’off-policy’ problem. Hence developing stable learning algorithms when used with function approximation is an important problem.
We present in this thesis a variant of Q-learning with linear function approximation that is based on two-timescale stochastic approximation. The Q-value parameters for a given policy in our algorithm are updated on the slower timescale while the policy parameters themselves are updated on the faster scale. We perform a gradient search in the space of policy parameters. Since the objective function and hence the gradient are not analytically known, we employ the efficient one-simulation simultaneous perturbation stochastic approximation(SPSA) gradient estimates that employ Hadamard matrix based deterministic perturbations. Our algorithm has the advantage that, unlike Q-learning, it does not suffer from high oscillations due to the off-policy problem when using function approximators. Whereas it is difficult to prove convergence of regular Q-learning with linear function approximation because of the off-policy problem, we prove that our algorithm which is on-policy is convergent. Numerical results on a multi-stage stochastic shortest path problem show that our algorithm exhibits significantly better performance and is more robust as compared to Q-learning. Future work would be to compare it with other policy-based reinforcement learning algorithms.
Finally, we develop an online actor-critic reinforcement learning algorithm with function approximation for a problem of control under inequality constraints. We consider the long-run average cost Markov decision process(MDP) framework in which both the objective and the constraint functions are suitable policy-dependent long-run averages of certain sample path functions. The Lagrange multiplier method is used to handle the inequality constraints. We prove the asymptotic almost sure convergence of our algorithm to a locally optimal solution. We also provide the results of numerical experiments on a problem of routing in a multistage queueing network with constraints on long-run average queue lengths. We observe that our algorithm exhibits good performance on this setting and converges to a feasible point.
40
Lakshmanan, K. "Online Learning and Simulation Based Algorithms for Stochastic Optimization." Thesis, 2012. http://hdl.handle.net/2005/3245.
Full textAbstract:
In many optimization problems, the relationship between the objective and parameters is not known. The objective function itself may be stochastic such as a long-run average over some random cost samples. In such cases finding the gradient of the objective is not possible. It is in this setting that stochastic approximation algorithms are used. These algorithms use some estimates of the gradient and are stochastic in nature. Amongst gradient estimation techniques, Simultaneous Perturbation Stochastic Approximation (SPSA) and Smoothed Functional(SF) scheme are widely used. In this thesis we have proposed a novel multi-time scale quasi-Newton based smoothed functional (QN-SF) algorithm for unconstrained as well as constrained optimization. The algorithm uses the smoothed functional scheme for estimating the gradient and the quasi-Newton method to solve the optimization problem. The algorithm is shown to converge with probability one.
We have also provided here experimental results on the problem of optimal routing in a multi-stage network of queues. Policies like Join the Shortest Queue or Least Work Left assume knowledge of the queue length values that can change rapidly or hard to estimate. If the only information available is the expected end-to-end delay as with our case, such policies cannot be used. The QN-SF based probabilistic routing algorithm uses only the total end-to-end delay for tuning the probabilities. We observe from the experiments that the QN-SF algorithm has better performance than the gradient and Jacobi versions of Newton based smoothed functional algorithms. Next we consider constrained routing in a similar queueing network. We extend the QN-SF algorithm to this case. We study the convergence behavior of the algorithm and observe that the constraints are satisfied at the point of convergence. We provide experimental results for the constrained routing setup as well.
Next we study reinforcement learning algorithms which are useful for solving Markov Decision Process(MDP) when the precise information on transition probabilities is not known. When the state, and action sets are very large, it is not possible to store all the state-action tuples. In such cases, function approximators like neural networks have been used. The popular Q-learning algorithm is known to diverge when used with linear function approximation due to the ’off-policy’ problem. Hence developing stable learning algorithms when used with function approximation is an important problem.
We present in this thesis a variant of Q-learning with linear function approximation that is based on two-timescale stochastic approximation. The Q-value parameters for a given policy in our algorithm are updated on the slower timescale while the policy parameters themselves are updated on the faster scale. We perform a gradient search in the space of policy parameters. Since the objective function and hence the gradient are not analytically known, we employ the efficient one-simulation simultaneous perturbation stochastic approximation(SPSA) gradient estimates that employ Hadamard matrix based deterministic perturbations. Our algorithm has the advantage that, unlike Q-learning, it does not suffer from high oscillations due to the off-policy problem when using function approximators. Whereas it is difficult to prove convergence of regular Q-learning with linear function approximation because of the off-policy problem, we prove that our algorithm which is on-policy is convergent. Numerical results on a multi-stage stochastic shortest path problem show that our algorithm exhibits significantly better performance and is more robust as compared to Q-learning. Future work would be to compare it with other policy-based reinforcement learning algorithms.
Finally, we develop an online actor-critic reinforcement learning algorithm with function approximation for a problem of control under inequality constraints. We consider the long-run average cost Markov decision process(MDP) framework in which both the objective and the constraint functions are suitable policy-dependent long-run averages of certain sample path functions. The Lagrange multiplier method is used to handle the inequality constraints. We prove the asymptotic almost sure convergence of our algorithm to a locally optimal solution. We also provide the results of numerical experiments on a problem of routing in a multistage queueing network with constraints on long-run average queue lengths. We observe that our algorithm exhibits good performance on this setting and converges to a feasible point.
41
Chang, Horng-Ying, and 張弘穎. "New results on fuzzy perceptron learning algorithm." Thesis, 1996. http://ndltd.ncl.edu.tw/handle/45155634444756303419.
Full textAbstract:
碩士國立交通大學
控制工程系
84
This thesis modifies a learning algorithm of fuzzy perceptron neural networksfor classifier that utilize expert knowledge represented by fuzzy if-then rulesas well as numerical data. We extend the conventional linear perceptron networkto a second order one, which can provide much more flexibility for discriminantfunction. In order to handle linguistic variables in neural networks, level setsof fuzzy set theory are incorporated into perceptron neural learning. At differentlevels of the input fuzzy number, the fuzzy perceptron algorithm is derived fromthe fuzzy output function and the corresponding nonfuzzy target output that indicatesthe correct class of the fuzzy input vector. The vertex method is borrowed andmodifies to obtain the extreme point of the fuzzy output function which can greatlyreduce the computational complexity and hence the time required for perceptronlearning algorithm. Moreover, the pocket algorithm is modified to our fuzzy perceptronlearning scheme, called fuzzy pocket algorithm, to solve the nonseparability problem,such as overlapping fuzzy inputs. Intensive computer simulations demonstrate theeffect of the modified algorithm, which solve the inaccuracy and speed problemsencountered in the Fuzzy BP algorithm of Tanaka.
42
Chen, Wei-Chou, and 陳偉洲. "Ontology-based Automatic Learning Objects Classification Algorithm." Thesis, 2006. http://ndltd.ncl.edu.tw/handle/57781709258226865423.
Full textAbstract:
碩士國立成功大學
工程科學系碩博士班
94
In the age of information explosion, the number of information people have to digest or deal with is over the edge of their tolerance. To hand over some manpower-consuming tasks to computers is one of the goals people pursue, which is also quite true in e-learning paradigm. With the standard convergence in the e-learning, most of the learning contents and learning objects are described by learning object metadata (LOM) that IEEE formulating. People can search easily from internet repositories and fetch many learning objects with standard LOMs. These learning objects can then be recombined and reused in different occasions. Therefore, if an automatic method for learning objects classification is available which groups learning objects into appropriate assortment, the jobs of recombination and reusing can be done quickly. In data mining, while there are many techniques for automatic classification, they are not suitable for automatic learning object classifications. This study proposes an ontology-based automatic learning object clas¬si¬fi¬cation algorithm. This algorithm focuses on analyzing the character¬istics of learning object metadata (LOM) and retrieving terms from LOM which help on classification. The power of the automatic classification of the algorithm comes from an ontology that domain expert constructed to guide the process of classification automatically.
43
Zhang, Hong-Ying, and 張弘穎. "NEW RESULTS ON FUZZY PERCEPTRON LEARNING ALGORITHM." Thesis, 1996. http://ndltd.ncl.edu.tw/handle/04192865218654218245.
Full text
44
JI, CHUN-GUAN, and 紀春全. "An incremental algorithm for learning from examples." Thesis, 1989. http://ndltd.ncl.edu.tw/handle/56274601760743080659.
Full text
45
Shih, Chie-Huai, and 施智懷. "Learning a Hidden Graph with Adaptive Algorithm." Thesis, 2008. http://ndltd.ncl.edu.tw/handle/87603380102822402500.
Full textAbstract:
碩士國立交通大學
應用數學系所
96
We consider the problem of learning a hidden graph using edge-detecting queries in a model where the only allowed operation is to query whether a set of vertices induces an edge of the hidden graph or not. Grebinski and Kucherov [5] give a deterministic adaptive algorithm for learning Hamiltonian cycles using Ο(log n) queries. Beigel et al.[4] describe an 8-round deterministic algorithm for learning matchings using Ο(log n) queries, which has direct application in genome sequencing projects. Angluin and Chen [2] use at most 12m(log n) queries in their algorithm for learning a general graph. In this thesis we present an adaptive algorithm that learns a general graph with n vertices and m edges using at most (2log n + 9)m queries.
46
Yu-HsuanHuang and 黃裕軒. "Deep Learning Applied to Speech Enhancement Algorithm." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/q2wt6u.
Full text
47
Zhuang, Ting-Wei, and 莊定為. "A NOVEL LEARNING-BASED LIDAR LOCALIZATION ALGORITHM." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/cvh79x.
Full textAbstract:
碩士國立臺灣科技大學
資訊工程系
107
Self-driving systems need to be able to localize its position with high accuracy regardless of whether it is during the day or night. This means because of the sensitivity to the lighting conditions, we cannot rely on ordinary cameras to sense the surrounding environment. A solution to replace the images is to use light detection and ranging sensor (LiDAR) to generate a three-dimensional point cloud of each point representing the distance to the sensor. In this paper, we propose a novel method for LiDAR localization using the three-dimensional point clouds generated by the LiDAR, a pre-build map, and a predicted pose as inputs and achieves centimeter-level localization accuracy. Our approach first selects a certain number of the online point cloud as key points. We then extract learned features from convolutional neural networks in order to train these neural networks to localize lidar. Our proposed method achieved significant improvements in terms of speed over prior state-of-the-art methods.
48
Lin, Yu-Shiou, and 林煜修. "Budgeted Algorithm for Linearized Confidence-Weighted Learning." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/wenc39.
Full textAbstract:
碩士國立交通大學
數據科學與工程研究所
107
This paper presents a novel algorithm for performing linearized confidence-weighted (LCW) learning on a fixed budget. LCW learning has been applied to solve online classification problems in recent years. To make better classification performance, it is common to combine with kernel functions through the kernel trick. However, the trick makes the LCW learning vulnerable to the curse of kernelization that causes unlimited growth in memory usage and run-time. To address this issue, we first re-interpret the LCW learning by using a resource perspective deeming every instance as a potential resource to exploit. Based on the perspective, we then propose a budgeted algorithm that approximates the LCW learning under a finite constraint on the number of available resources. The proposed algorithm enjoys finite complexities of time and space and thus is able to break the curse. Experiments on several open datasets show that the proposed algorithm approximates the LCW learning well and is competitive to state-of-the-art budgeted algorithms.
49
Joshi, Varad Vidyadhar. "Expert-gate algorithm." Thesis, 1992. http://hdl.handle.net/1957/36248.
Full textAbstract:
The goal of Inductive Learning is to produce general rules from a set of seen
examples, which can then be applied to other unseen examples. ID3 is an inductive
learning algorithm that can be used for the classification task. The input to the
algorithm is a set of tuples of description and class. The ID3 algorithm learns
a decision tree from these input examples, which can then be used for classifying
unseen examples given their descriptions. ID3 faces a problem called the replication
problem.
An algorithm called the Expert-Gate algorithm is presented in this thesis. The
aim of the algorithm is to tackle the replication problem. We discuss the various
issues involved with each step of the algorithm and present results corroborating
our choices. The algorithm was tested on various artificially created problems as
well as on a real life problem. The performance of the algorithm was compared
with that of Fringe.
The algorithm was found to give excellent results on the artificially created
problems. The Expert-Gate algorithm gave satisfactory results on the NETtalk
problem. Overall, we believe the algorithm is a good candidate for testing on other
real life domains.Graduation date: 1993
50
Huang, Sheng-Bo, and 黃聖博. "Learning Recommendation System based on Micro-Learning Materials and Data Mining Algorithm." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/bhb7v9.
Full textAbstract:
碩士南臺科技大學
資訊管理系
105
Information overload is a most encountered problem in learning, especially for college students. Learners need to complete a lot of compulsory and elective subjects within limited time. Besides, the content of those subjects keeps diversity due to improvement of knowledge and technology. Therefore, it is difficult to learn full knowledge only through textbook materials. Learners needs to seek extra learning materials via Internet. But it contains a lot amount of information which makes learning or reading time-consuming. And it also causes information overload issue. Besides, every learner has different learning ability and prerequisite knowledge due to individual difference situation. An individual difference situation means that learners have same learning materials and tutors but with different learning outcomes. In order to improve information overload issue and individual differences situation, this research proposes a learning recommendation system based on micro-learning materials and data mining algorithm. The system utilizes automatic summarization technology and personal recommendation mechanism to overcome the issues mentioned above. The experiment reveals that the automatic summarization technology produces highly readable content for readers. And the experiment also reveals that most learners are able to get different recommended learning path according to their learning history from the proposed system.